Two big additions in this release. Asset Management gives you a built-in media library for uploading, organizing, and referencing files directly inside your AI chat. Skills let you teach the AI custom workflows and design patterns that it follows automatically — including multi-page edits dispatched in parallel.
Asset Management
Every Capuzzella site now has a dedicated /assets page where you can upload
images, videos, and documents, sort them, filter them, and delete the ones you no longer
need. Files are served from your project's public directory the moment they're uploaded,
so they're available everywhere on your site instantly.
The most useful part is the AI integration. Type @ in the Editor Panel and
your uploaded assets show up alongside components and pages in the autocomplete. Pick
one and the AI receives its URL as part of the prompt, so it can drop the right
<img> or <video> tag into your page without you
having to remember any paths. Assets used on published pages are also protected from
deletion, so you can't accidentally break a live image link.
Read the full Asset Management guide →
Skills
Skills are small Markdown files that teach your AI assistant how to handle specific tasks
or follow your design conventions. They live in Settings → Skills,
and you reference one in chat by typing @skill-name. Each skill is either
essential (loaded on every request) or available (loaded on demand
when the AI needs it), so you can keep the context window lean while still having deep
expertise on tap.
This release also introduces multi-page edits. A skill can now instruct the AI to edit several pages at once — for example, a "publish blog post" skill that formats the current page and adds a new entry to the blog index in a single request. Behind the scenes, the planner dispatches one worker per page in parallel, so the wall-clock time barely changes whether you're editing one page or ten.
Try It Out
Both features ship today on every Capuzzella instance. Upload your first asset, write your first skill, and watch the AI chain them together into the kind of workflow that used to live in your head.